Pattern and colour grading and crack detection on marble slabs
A marble slab from the polishing line is imaged for colour and vein pattern under diffuse light. A deep learning model assigns its selection class; a second low-angle image is checked for cracks and a vacuum gripper moves the slab to the corresponding stack.
Selection class and defect position appear on the operator display; line control sends the target stack to the vacuum stacker. A cracked slab is placed on a separate stack.
- 01 Infeed
- 02 Acquire
- 03 Analyse
- 04 Decide
The slab enters the enclosure on the roller conveyor.
- Selection class
- —
- Crack length
- —
Scenes are illustrative simulations; they are not customer references or performance guarantees.
Need
Natural veins and fossils may resemble cracks; reflections on the polished surface alter the base colour. A single lighting geometry is unsuitable for both colour and fine cracks.
Solution approach
Diffuse panels in the enclosure provide an overhead colour image of the full slab; a deep learning model assigns a class from pattern similarity and base colour. A low-angle bar light then activates; crack and filler regions are assessed in the second image.
Process steps
- 01
Infeed
The slab enters the enclosure on the roller conveyor.
- 02
Acquire
A diffuse colour image is followed by a low-angle surface image.
- 03
Analyse
A pattern class is assigned; the crack region is marked in the image.
- 04
Decide
The vacuum gantry moves the slab to the stack selected by the verdict.
Related applications
Let us assess your application together
We test your sample in our laboratory and report on resolution, optics and lighting selection.